The next-generation sequencing revolution is enabling unprecedented access to causal genes underlying a variety of disease conditions. This information promises to lead to more effective and increasingly personalized therapeutics as new disease mechanisms are characterized and target genes are identified. A critical bottleneck in leveraging this information to the point of defining new treatments, however, is the development of safe and effective therapeutics, which are often small molecules that bind the protein target of interest. Even with a well-defined target, development of small molecule probes is expensive and inefficient, which is why it can take years or even decades of drug development from discovery of the disease mechanism to an FDA approved drug. The proposed research addresses this bottleneck with the long-term goal of rapidly characterizing novel compounds' modes of action to build a comprehensive small molecule library targeting a significant fraction of the human genome. The specific objective of this application is to develop key computational infrastructure for high-throughput chemical genomics approaches, which leverage model organism mutant libraries as a diagnostic for compound target discovery. This objective will be accomplished through three specific aims: (1) the development of an experimental pipeline and computational infrastructure for chemical genetic interaction mapping in S. cerevisiae, S. pombe, and E. coli and application of the approach to large libraries of natural products or synthetic compound libraries, (2) the development of methods for combining chemical-genetic and genetic interactions to predict mode-of-action for large compound libraries, and (3) the development and experimental validation of predictive models for compound synergy. The proposed research is innovative because it closely integrates computational approaches leveraging the structure of genetic interaction networks with optimization of a powerful experimental assay. Furthermore, it challenges the current paradigm of target-centric therapeutic development as well as the notion of an inherent tradeoff in compound screening throughput when chemical genomic approaches are used. The proposed work will demonstrate that chemical genomics can be scaled to accommodate the largest of chemical libraries while providing an unbiased strategy for identifying novel modes of action. Other expected outcomes include (1) the discovery of hundreds of new small molecule probes with precise modes of action, (2) methods for integrating genome-scale data across species to improve the relevance of model organism chemical- genetic data to human health, (3) fundamental characterization of how the diversity of natural products interacts with eukaryotic cells on a global scale, and (4) mechanistic understanding, predictive models, as well as several novel discoveries of compound combinations that act synergistically.